You seem to come across the term ‘Deep Learning’ everywhere these days. It’s all pervasive and seems to be at the heart of all AI related research. It has even spawned new and never-thought-of-before innovations!
But how can you learn it? There are way too many resources out there, spread in a very unstructured and not a very beginner friendly manner. You complete a course on one platform, move to another course on a different platform, and so on. You learn, but not in any logical or sequential manner. That’s a bad idea.
We have put together a comprehensive learning path for any person wanting to get into the field of deep learning. This path contains plenty of resources, links, ideas and suggestions to get you on your way! I encourage you to check out the full article here, which contains these resources.
The below infographic is a very handy resource every aspiring and even established data science professional should keep handy with them at all times. You can download a high resolution PDF (just 1.5MB) from this link. I highly recommend doing this and taking a printout of the infographic – it’s a very handy pocket guide.
This is your ticket to deep learning – use it wisely!
Lorem ipsum dolor sit amet, consectetur adipiscing elit,
I notice that most of deep learning projects are for audio / video / images / text. What about the traditional use cases like classification / regression. I am curious why that is not popular for deep learning? Thank you
I was recently at a talk by JJ Allaire, the author of 'Deep Learning with R' and this is what he says. Computer Vision, Reinforcement Learning (like playing Go) and Speech and Natural Language Processing (to some extent) are the areas where deep learning has been proved to be better than traditional models, so far. For other problems, it is not clear if deep learning models outperforms traditional machine learning models and is additionally more resource-intensive. As a side note, I would love to hear of other use cases where people have found deep learning models to outperform traditional models.